GoVector: An I/O-Efficient Caching Strategy for High-Dimensional Vector Nearest Neighbor Search

2025-08-21

AI Summary

To address the excessive query latency—attributable to I/O-bound overhead (over 90% of total latency)—in disk-resident graph indexes for high-dimensional Approximate Nearest Neighbor Search (ANNS), this paper proposes a hybrid caching strategy integrating static preloading with dynamic adaptive caching. It further introduces a novel similarity-based disk-node reordering technique to enhance vector-space locality, coupled with page-level data layout optimization to improve cache hit rates. Experimental results demonstrate that, at 90% recall, the proposed method reduces average I/O operations by 46% compared to the state-of-the-art systems, increases query throughput by 1.73×, and decreases end-to-end latency by 42%, thereby significantly alleviating the I/O bottleneck inherent in disk-based graph indexing.

Technology Category

• Search and Optimization: Distributed Search
• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Graph-based high-dimensional vector indices have become a mainstream solution for large-scale approximate nearest neighbor search (ANNS). However, their substantial memory footprint often requires storage on secondary devices, where frequent on-demand loading of graph and vector data leads to I/O becoming the dominant bottleneck, accounting for over 90% of query latency. Existing static caching strategies mitigate this issue only in the initial navigation phase by preloading entry points and multi-hop neighbors, but they fail in the second phase where query-dependent nodes must be dynamically accessed to achieve high recall. We propose GoVector, an I/O-efficient caching strategy tailored for disk-based graph indices. GoVector combines (1) a static cache that stores entry points and frequently accessed neighbors, and (2) a dynamic cache that adaptively captures nodes with high spatial locality during the second search phase. To further align storage layout with similarity-driven search patterns, GoVector reorders nodes on disk so that similar vectors are colocated on the same or adjacent pages, thereby improving locality and reducing I/O overhead. Extensive experiments on multiple public datasets show that GoVector achieves substantial performance improvements. At 90% recall, it reduces I/O operations by 46% on average, increases query throughput by 1.73x, and lowers query latency by 42% compared to state-of-the-art disk-based graph indexing systems.

Problem

• Reducing I/O bottlenecks in disk-based graph indices for ANNS
• Improving cache efficiency for dynamic query-dependent node access
• Optimizing storage layout to enhance spatial locality and reduce I/O

Innovation

• Static and dynamic caching for graph indices
• Disk node reordering for spatial locality
• Similarity-driven storage layout optimization